Chondrules

Earth and Planetary Science Letters 224 (2004) 1 – 17 www.elsevier.com/locate/epsl

Frontiers

Chondrules B. Zanda * Laboratoire d’Etudes de la Matie`re extraterrestre MNHN-UMS2679 CNRS 61, rue Buffon 75005- Paris, France Department of Geological Sciences, Rutgers University, P.O. Box 1179, Piscataway, NJ 08855-1179, USA Received 17 June 2003; received in revised form 5 May 2004; accepted 5 May 2004

Abstract Chondrules are submillimeter spheres that constitute up to 80% of the volume of the most primitive meteorites. That they result from the solidification of a melt in low-gravity in the early solar system has been known for nearly two centuries, but the conditions of their formation and their significance still elude our understanding. It has been variously proposed that they both predate and postdate planet formation and some have suggested that they are the product of the planet-forming process itself. There is mounting evidence that rather than resulting from a trivial event (or series of events) which melted only a small fraction of solids in the disk, chondrule formation significantly transformed the original material present in the early solar system and contributed to the chemical and isotopic compositions of the first planets. The only meteorites that preserved the chemical composition and isotopic signatures of the earliest solar system solids are the CI chondrites that contain no preserved chondrules and probably had very few, if any. All other chondrites have experienced various levels of metal/silicate and refractory/volatile fractionation that may have resulted from chondrule formation, although a number of researchers argue that these fractionations existed before chondrules and probably resulted from nebular-wide condensation. The current most popular mechanisms for forming chondrules in a nebular setting are radiation emitted by the protosun in the X-wind setting or shock waves propagated in the protoplanetary disk. In the latter case, chondrule formation may have contributed to the first stages of accretion, which would have helped preserve the chemical complementarity between chondrules and matrix. It is important that the chemical and isotopic properties, and even the petrology, of chondrules be reassessed in order to allow the development of chondrule formation models that better fit these constraints. D 2004 Elsevier B.V. All rights reserved. Keywords: chondrules; chondrites; formation of the solar system; solar nebula; early chemical fractionations

1. Introduction 1.1. Chondrules and chondrites The most common kind of meteorites that have been observed to fall on the Earth are chondrites. These are * Laboratoire d’Etudes de la Matie`re extraterrestre MNHNUMS2679 CNRS 61, rue Buffon 75005- Paris, France. Tel.: +33-14079-3542; fax: +33-1-4079-5772. E-mail address: [email protected] (B. Zanda). 0012-821X/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2004.05.005

like no rocks formed on Earth and there now exists a sizeable body of data to indicate that they formed from dust and debris from the solar nebula that escaped being incorporated into planets. As such, they provide unique insights into processes operating in the circumstellar disk from which the planets formed at the very start of the solar system. Chondrites are so named because they usually contain large amounts of small (at most millimeter-sized) spherules called chondrules. Although these are the most abundant objects in chondrites, the origin of chondrules is extremely

2

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17 Jargon Box

Calcium – Aluminum-rich inclusions (CAIs) CAIs are particles up to a centimeter in size found in chondrites. They are made of minerals containing refractory elements such as Ca, Al and Ti. CAIs are the oldest objects of our solar system which they allowed us to date. They are considered to be either the first objects condensed within the solar system or the only residue of the wholesale evaporation which took place during its formation. A fraction of the CAIs have been melted and their relationship with chondrules is unclear. They differ in compositions but a few intermediate objects exist. They also appear older than chondrules. It is still unknown whether the event(s) that melted some of them is related to chondrule formation. Chondrites The meteorites that hit the Earth fall into two broad categories. ‘‘Differentiated meteorites’’ come from asteroids which were melted (like the larger planets) and they sample different layers from these bodies: the eucrites (most of which are basalts) are crust samples while the iron meteorites are core samples. Primitive meteorites, on the other hand, come from bodies which were never melted after their formation in the early solar system and they have preserved characteristics acquired in the solar accretion disk. They are called ‘‘chondrites’’ because their most striking feature is that they are made of up to 80% chondrules. Three main categories of chondrites were recognized early on: the ordinary (OCs, most common), enstatite (ECs, so reduced that some lithophile elements such as Mn or Ca are found in sulfides) and carbonaceous chondrites (CCs), in which organic matter was detected and a large fraction of which are oxidized (containing no metal but magnetite). We now know that these three categories encompass different domains in the oxygen isotopes diagram shown in Fig. 9 (above, on and below the TFL). These three categories are also divided into subcategories: EH and EL for the ECs depending on their high (H) or low (L) content of Fe; H, L and LL for OCs depending both on their total Fe and their metal content (Fig. 8). A fourth group, apparently related to LLs but even more oxidized, also exists: Rumurutiites (Rs). The subcategories of OCs and even ECs are fairly closely related both in terms of mineralogy and of oxygen isotopes. The CCs are much more diverse and the most of the groups are named after a type meteorite: CI (Ivuna); CM (Murray); CV (Vigarano); CO (Ornans); CR (Renazzo); CK (Karoonda); CH (extremely High in metal). Chondrules Chondrules are submillimeter particles found in chondrites of which they can comprise most of the mass. Chondrules are often spherical and are believed to be derived from liquids crystallized in low gravity in the early solar system. Chondrules have a large variety of textures and compositions but they fall into two main chemical categories shown in Fig. 2 and defined in its caption.

controversial and has generated more debate than any other feature of meteorites. Their formation may play a major role in determining the composition of the inner planets (e.g., [1]). Now scientists are at the point of being able to propose new theories that are being stimulated by the exciting new information being obtained from astronomical observations of disks around young solar mass stars coupled with astrophysical modelling. 1.2. Formation of the Sun and solar system The formation of stars and planets is now better understood but the occurrence of widespread melt droplets, chondrules, is not a feature that has been predicted in such work. It is likely that the gravitational collapse of a dust-gas cloud was triggered by the blast from a supernova, which injected newly synthesized isotopes into the future solar system. The cloud material spiralled inwards, forming a Sun and circumstellar disk from which planetary material accreted. The inner regions of this protoplanetary disk, or solar nebula,

must have been hot enough to evaporate all solids, but we do not know how extensive or long-lived was the heating in this region. Material passing from the disk into the growing Sun would have encountered its magnetosphere and it has been proposed that it would have been divided into an inflow and an outflow known as the X-wind [2]. Much mass would have been lost in bipolar outflows but some particles may have decoupled from the X-wind and fallen back into the disk. Whether or not the X-wind model is correct, there were numerous energetic processes within the disk, as well as at its inner edge, some of which may have led to the formation of chondrules by melting solids or by condensing vaporized solids. 1.3. Controversies about chondrules Linking chondrules with such new theories for disk behavior based on observational astronomy and astrophysics may well be the best way to determine their origin. Three years ago, at the LPS Conference in Houston, a prominent meteoriticist gave a highly

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

controversial plenary lecture [3], later featured in Science [4], in which he expressed his doubts as to whether continuing to accumulate data on chondrules would ever lead to a better understanding of their origin. This is because of the lack of a grand theory into which these data could be fitted. Indeed, the diversity of models that have been proposed to account for chondrule formation (e.g., Fig. 1) is a testimony as to our present lack of understanding of their origin and our need for a framework for understanding the physical and chemical properties of the first solar system bodies. However, as pointed out by Hewins [5], the lack of agreement between meteoriticists about the detailed interpretation of chondrule petrology and geochemistry is part of the reason for our inability to develop such a unifying theory. As suggested by Wood [6], it is important to reevaluate the data and the paradigms on chondrule formation, taking into account more recent developments. However, contrary to Wood [3] and Kerr [4], a strong case can be made that the acquisition of new data taking full advantage of recently developed techniques remains essential, especially on key issues such as formation ages, irradiation signatures and exchanges with the surrounding gas.

3

Progress will also come from the refinement of astrophysical models based on new astronomical observations that are more closely tailored to the detailed physical and chemical properties of chondrules. The present paper briefly reviews the main characteristics of chondrules and the relevant properties of chondrites, the leading arguments concerning the origin of chondrules and the nature of their precursor material, and the approaches that presently appear the most promising to unravel this mystery.

2. Key properties of chondrules 2.1. Petrography and chemistry of chondrules Chondrules are small particles of silicate material that experienced melting before their incorporation into chondritic parent bodies. Particles with different compositions such as Calcium –Aluminum-rich Inclusions (CAIs) and basaltic fragments are not considered chondrules [5]. Ideal chondrules are spherical as they solidified from liquid droplets, but most chondrules are not spheres; that is, a significant fraction is

Fig. 1. This cartoon was sketched by the astrophysicist P. Cassen at the ‘‘Chondrule Conference’’ in 1994 as an up-to-date summary of chondrule formation models. The profusion of energy sources makes it a wonder that any material in the disk escaped melting (e.g., the chondritic matrix and the fluffy CAIs). Note that the details of the various models have evolved significantly since.

4

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

present as fragments (Fig. 2), some are molded around one another (Fig. 3a) while others were insufficiently melted for surface tension to make them round (Fig. 3b). Figs. 2 and 3 illustrate some of the key properties of chondrules that any model of chondrule formation must account for the variability of their shapes, textures and compositions, starting with the existence of two main chemical classes, one of which is volatile-rich. Another key point is the evidence for extended or successive episodes of heating in the form of melted rims around some chondrules and relicts from earlier generations of objects. These relicts are usually found as grains (Fig. 4), but chondrules from CO and CV chondrites may also contain isotopic geochemical signatures derived from CAIs [7] or amoeboid olivine aggregates [8]. 2.2. Isotopic data More evidence pertaining to the chondrule formation mechanism stems from isotopic measurements in chondrules. One intriguing problem is that no or only

Fig. 2. This reflected light image of a primitive ordinary chondrite shows that it looks like a ‘‘sediment’’ of subspherical particles, chondrules, embedded in a fine-grained matrix loaded with chondrule debris. Chondrules can have highly variable sizes and textures (e.g., [5]), but the main distinction is chemical: volatiledepleted type I chondrules are reduced and often metal-bearing (lower-right), whereas metal is absent from volatile-rich type II chondrules in which Fe is oxidized within the silicates or in sulfides (left). Type II chondrules tend to be coarser-grained than type Is and are often also larger. Chondrules of the two types also have variable Mg/Si contents: some of them are dominantly olivine-rich while others are pyroxene-rich. (Field: 2.3
3 mm).

Fig. 3. In some primitive ordinary chondrites, chondrules can be molded around one another (a) indicating that they must have come together while the deformed chondrule was still above its glass transition temperature. In primitive carbonaceous chondrites, chondrules are often contorted (b) and appear like an aggregation of several subunits together with some dust. Some of the subunits and the chondrule as a whole were insufficiently melted to adopt an overall droplet shape. [Field width: 2 mm (a); 3 mm (b)]. On these BSE images, reduced silicates (type I chondrules) appear dark while oxidized material (type II chondrules and matrix) appears light grey, and metal and sulfides are white. These pictures give an idea of the differences between ordinary and carbonaceous chondrites in which chondrules may have (slightly?) different origins. Carbonaceous chondrites have an abundant matrix and mostly type I chondrules, whereas type II chondrules are largely dominant in ordinary chondrites.

very minor isotopic mass fractionations have been found in chondrules so far, even in volatile-depleted ones (e.g., [9,10]). Yet Rayleigh distillation would be expected if chondrules experienced free evaporation. As will be discussed below, this could either indicate that chondrule heating was extremely brief or that the

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

Fig. 4. These two chondrules from a primitive ordinary chondrite contain a relict from a small broken chondrule (a) and several grains from an FeO-rich chondrule which became ‘‘dusty’’ with minuscule beads of Fe produced by the reduction of their FeO when they were included in an FeO-poor chondrule. The dusty grains are visible both in transmitted light (b) and in reflected light (c) where the metal beads on the polished surface can be distinguished. [Field width: 2.2 mm (a); 0.8 mm (b and c)].

environment in which they melted significantly differed from the canonical solar nebula, e.g., in the partial pressures of the lithophile elements. Another key question is that of the age of chondrules. Based on the decay of extinct radionuclides such as 26Al, 53Mn and 129I, Swindle et al. [11] suggested that the formation of chondrules took place up to several million years after that of CAIs, the earliest solar system objects. The chronological significance of the data for each of these nuclide systems has been the subject of debate. Pb– Pb isochron ages of individual chondrules from a CR chondrite have recently been measured by Amelin et al. [12] and compared to similar ages obtained in CAIs of a CV chondrite. The chondrules were formed at 4564.7 F 0.6 Ma, 2.5 F 1.2 My after the refractory inclusions (an age difference in agreement with that based on 26Al). This

5

indicates that chondrules could not have been formed in the earliest stage of infall and rapid accretion of the solar nebula as advocated by Wood [6]. Moreover, some asteroids and planetary embryos may have been already accreted by the time chondrules were melted and incorporated in chondritic asteroids [13] as recent calculations show planetary embryos forming in much less than 1 million years [14]. Heterogeneities in the spallogenic nuclides 11B and 7 Li in chondrules were recently reported by Robert and Chaussidon [15] and were interpreted to result from the irradiation of chondrule precursor material by high energy (MeV) particles (mostly protons, but also alphas and possibly 3He nuclei) emitted by the Sun during its T-Tauri phase. Assuming these results are confirmed, they would constitute the first unequivocal proof of the early irradiation of solar system material because, unlike other short-lived nuclides such as 26Al and 53 Mn, these nuclides cannot be formed in supernovae or other stars. The presence of these nuclides offers little additional constraint on the age of chondrules (since the T-Tauri phase of the Sun may have lasted up to a few million years). It would, however, place some constraints on the location where chondrules formed if the irradiating particles originated in the Sun (rather than in Galactic Cosmic Rays as suggested by Desch et al. [16]). MeV particles could only have penetrated as far as 3 AU, the present location of the asteroid belt, if the nebula was ‘‘thin’’ (i.e., low density). Chondrules, however, are unlikely to have formed at the ultralow pressures (below 10 8 atm) prevailing in a thin nebula. If the nebula was ‘‘thick’’, and chondrules were irradiated by the Sun, they must have been formed closer to the Sun and then transported out to the asteroid belt. 2.3. Experimental simulations Experimental petrology also yields information on the chondrule forming process. Textures of chondrules were reproduced successfully about 15 years ago (e.g., [17]) allowing us to derive estimates of peak temperatures and cooling rates. Most chondrules appear to have been heated to about 1500 –1600 jC [18] although their liquidus temperatures vary enormously (f 1200– 1900 jC). Their cooling rates ranged from about 10 to 1000 jC/h, much slower than the radiative cooling of isolated spherules into free space, and much faster than global nebular cooling [5]. This is

6

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

considered evidence that chondrules were produced in large quantities embedded in hot gas [19]. More recent experiments have focused on reproducing the chemical and isotopic properties of chondrules. Yu et al. [20] showed that the oxygen isotopes of a chondrule could result from an exchange between its precursor material and the ambient gas, while in [18,21], they demonstrated that the volatile content of type II chondrules were incompatible with their being formed in the canonical solar nebula. On the other hand, Cohen et al. [22] showed that type I chondrules could have formed from the severe degassing of type II material and Cohen [23] established that such degassing need not result in Rayleigh distillation, as the presence of a volatile-rich atmosphere in the chondrule forming region would have been sufficient to prevent such fractionation. Suppression of isotopic fractionation by exchange during evaporation has also been modelled by Alexander [24].

3. Relevant properties of chondrites

Fig. 5. Transmitted light view of PTS from a CO chondrite and a CV chondrite at the same scale (field 7
4.6 mm). Particles that make up chondrites (chondrule, chondrule fragments, metal grains and CAIs) appear to have a very restricted size distribution within a given chondrite class but to differ very significantly from class to class, ranging from an average of 20 Am in CHs and 150 Am in COs up to about 1 mm in CVs. The abundance of matrix also differs significantly ranging from 5% in CHs up to >99% in CIs. In Cos, the matrix amounts on the average to f 34% and in CVs to f 40% [66].

3.1. Petrographic evidence It would be hard to discuss the origin and significance of chondrules without taking into account some of the key properties of chondrites. As can be seen in (Figs. 2, 3 and 5), chondrites are an assemblage of chondrules and interstitial matrix in highly variable proportions. While chondrules (and chondrule fragments) may constitute up to 80% of the volume of some ordinary chondrites (Fig. 3a), their abundances were probably lower in carbonaceous chondrites (Fig. 3b) and close to zero in CI chondrites. Their sizes (Fig. 5), compositions and textures also vary between chondrite classes. Ferroan and droplet shaped chondrules are much more frequent in ordinary chondrites than in carbonaceous chondrites (Fig. 3). The composition and nature of the fine-grained matrix also varies within an individual chondrite and from one chondrite to another [25]. While all chondrites experienced some degree of aqueous alteration and/or metamorphism, chondrites that have the most matrix appear to be most altered to phyllosilicates, whereas the least altered matrix consists of a complex mixture of material from different sources, including presolar grains, chondrule fragments and condensate material [25].

Apart from the fact that they contain chondrules, the most striking characteristics of chondrites are that their compositions are close to that of the Sun and that metallic minerals tend to be associated with chondrules in carbonaceous chondrites (Fig. 2 and 3b), and are interstitial between chondrules in ordinary chondrites. 3.2. CI chondrites: the solar system reference The chondrites that most closely mimic the composition of the solar system are the CI chondrites (Fig. 6), which have the peculiarity of containing no or no surviving chondrule and consist almost entirely of highly altered matrix. CI chondrites contain no metal and are so oxidized that magnetite is present. They contain abundant presolar grains and their Cr [27] and Mo [28] were both discovered to consist of two complementary isotopically anomalous components, which appear also to be present in the matrix of other carbonaceous chondrites [27]. It has been argued that the extensive parent-body alteration undergone by CI chondrites was responsible for the destruction of the chondrules and CAIs they

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

7

material, like primitive chondrite matrix, is unlikely to have ever been melted or vaporized and recondensed in the nebula. This explains why CIs are the only meteorites that retained a solar composition and why they are so extensively altered compared to other chondrites: a greater amount of ice would have survived in a chondrule-free region to alter the silicate minerals. This idea receives some support from the detection by [32] in chondrule pyroxenes of water enriched in D up to three times above the values of the matrix and of the chondrule mesostasis. That is, the water in the mesostasis was probably introduced during low temperature alteration, but the water in the pyroxenes must have a different origin, implying that chondrules were formed in a water-rich environment or from a water-bearing material. Fig. 6. With the exception of the most volatile elements (H, rare gases, C and O), CI chondrite compositions exactly match those of the solar photosphere which is representative of the sun, i.e., 99.99% of the mass of the solar system. This agreement holds for major and trace elements (after [26]—note scale is exponents).

originally contained (e.g., [29]). The presence of relict crystals fallen from chondrules or CAIs [29,30] however does not prove that more than 1% of CI material was ever melted. On the contrary, the presolar grain concentrations, the heterogeneities in the Cr and Mo isotopes and the isotopic compositions of organic material [31] all indicate that the vast bulk of CI

3.3. Chemical fractionations among chondrite classes All other chondrite classes are chemically fractionated with respect to CIs (carbonaceous chondrites are shown in Fig. 7) and the variability of their Fe content and of its oxidation state has long been used as the classification reference (Fig. 8). As will be discussed below, chondrule formation involves both evaporation/ condensation and redox processes and has the ability to generate the metal that is absent from CIs. A comparison of the chemical and isotopic properties of the chondrule-bearing chondrites with those of CIs thus suggests that chondrule formation may have played an

Fig. 7. All chondrite classes (and the Earth’s mantle) exhibit chemical fractionations compared to CIs and the Sun. In carbonaceous chondrites, the more volatile elements are more fractionated, suggesting the fractionation resulted from a high temperature process involving evaporation and/or condensation (from [33]).

8

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

Fig. 8. Chondrite classes were originally defined based on their total Fe content and its oxidation state. In this modified version of the Urey and Craig diagram (after [26]), total Fe is constant along the diagonals, increasing from lower left to upper right. CIs have no metal although only EHs have more total Fe. As for their other chemical properties and the abundance of their chondrules, CMs are intermediate between CIs and the other carbonaceous chondrite groups, having less total Fe and a small amount of metal. EH chondrites are enriched in Fe and in volatiles in general with respect to CIs.

essential role in establishing those properties and generating the chondrite classes. This hypothesis was originally proposed for the volatile/refractory element fractionation as the ‘‘two-component’’ model of Larimer and Anders [34] and is still supported by the complementarity that apparently exists between chondrules and their surrounding matrix [35 – 37]. That is, when chondrules are more refractory, the matrix tends to be either more abundant or more volatile-enriched, allowing the bulk chondrite to have a bulk composition that is close to CI and significantly less fractionated than each of the individual constituents. 3.4. Oxygen isotopes The various chemical chondrite classes, the Earth – Moon system and the planet Mars all have distinct oxygen isotopic signatures that cannot be explained by mass dependent fractionation alone (Fig. 9). A mixing from at least two different reservoirs lying on a slope 1 line (i.e., containing different amounts of 16O) is required as well. Since their first discovery by Clayton

et al. [38] almost 30 years ago, the nature and significance of these distinct oxygen reservoirs have been a matter of a heated debate. Recent observations seem to shed new light on these issues and allow us to generate the various reservoirs from the mixing of a limited number of end-members. Following Clayton and Mayeda [39], Young and Russell [40] demonstrated that alteration of CAI minerals shifted their isotopic compositions towards heavier oxygen along massdependent fractionation lines whereas unaltered minerals fall along a slope 1 line passing through the ordinary chondrite groups (Fig. 9). This allows us to distinguish two types of vector on the oxygen isotopic diagram, related to different physical processes: massdependent fractionation towards the right, which is related to alteration (and may be proportional to the amount of matrix in the material [45]) and massindependent effects along the f 1 slope line ‘‘Y + R’’ which remains to be explained. Results by Luck et al. [41] however indicate that this mass-independent fractionation may derive from the process that fractionated volatiles from refractories in chondrites: (1) 16O excesses are coupled with 63Cu excesses both in carbonaceous and (along a distinct trend) in ordinary chondrites [41]; (2) there is a negative correlation between the 63Cu excesses of the various chondrite groups and their ratios of moderately volatile elements such as Mn [41], Na and K [46] relative to the refractory element Al. As discussed below, the process that fractionated volatiles from refractories could be closely related to CAI/chondrule formation since CAIs are enriched in 16O whereas the most volatile-rich chondrules are the most depleted. 3.5. Clasts in chondrites as a clue to their formation environment Other noteworthy properties of chondrites include the fact that they contain lithic clasts that might have originated from differentiated asteroidal bodies as advocated by Hutchison [47]. Together with the presence of numerous chondrule fragments, this has been taken as evidence for the occurrence of collisions and fragmentation before the final accretion of chondrites, which was therefore suggested to have taken place in a planetary environment [47]. This interpretation appears consistent with the metamorphic Pb– Pb age of the oldest dated chondrites (4.563 Ma [48])

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

9

Fig. 9. For convenience, the small variations of oxygen isotopic compositions are described in terms of their deviation in parts per thousand from a terrestrial standard, Standard Mean Ocean Water (SMOW): d17 O ¼ f½ð17 O=16 OÞsample =ð17 O=16 OÞSMOW   1g
103 d18 O ¼ f½ð18 O=16 OÞsample =ð18 O=16 OÞSMOW   1g
103 : Each chemical class of chondrites has its own oxygen signature (with some amount of overlap between classes). For simplicity, only the major carbonaceous chondrite classes are displayed here and EH and EL chondrites are grouped as ‘‘E’’. Two types of isotopic fractionations are involved: mass-dependent and mass-independent. Mass-dependent fractionations are usually the result of temperature, coordination or kinetic effects and move the samples along a slope 1/2 line. Except for some very special cases, all terrestrial samples are thus restricted to the 1/2 slope line labelled ‘‘terrestrial fractionation line’’ (TFL), and each chondrite class is characterized by its D17O (the excess of d17O relative to the terrestrial fractionation line). In meteorites, mass-independent effects are also present, the origin of which is still not properly understood. Most CAIs and chondrules from CV chondrites lie along a slope f 1 line known as ‘‘carbonaceous chondrite anhydrous mineral’’ (CCAM) which was suggested to result from the addition of pure 16O to these samples [38]. Ordinary chondrites chondrules all fall in the OC domain, i.e., above the TFL. A new understanding of this diagram is emerging. Following the suggestion of [39], Young and Russell [40] recently showed that unaltered minerals from CAIs lie on a slope 1 line (‘‘«Y + R»’’) going through the ordinary chondrite groups, whereas altered minerals are shifted to the right (along a slope 1/2 line). The D17O of the chondrite groups (which reflects their projection along the TFL onto the «Y + R» line) are related with their 63Cu excesses which are, in turn, correlated with their Mn/Al ratios [41]. Data sources: Clayton and Mayeda [42], Clayton [43], Kalleyman et al. [44] and Young and Russell [40].

which compares to that of eucrites (4.56 Ma [49]). This theory is, however, not widely accepted.

4. The origin of chondrules: their precursors and links with other chondritic components The question of the origin of chondrules is probably the most tantalizing one in meteoritics. As discussed in [3], we are still in search of a grand theory and an embarrassing number of open questions remain despite the wealth of accumulated data. These questions fall into two closely connecting broad categories: the nature of the material from which the chondrule liquids were generated and the physical mechanism responsible for the melting event. To discuss these issues, the basic observations presented above will be connected with additional recent significant observations.

4.1. The canonical view It is first necessary to summarize the ideas on chondrules that were most commonly accepted until recently, although some appear to have little other grounding than having been repeated over time, as pointed out by Wood [6]. In the canonical view, equilibrium condensation (resulting from the cooling of an initially entirely vaporized inner nebula of solar composition) generated small solid grains. These grains then aggregated together into small dustballs which were heated to high temperatures by an as yet unidentified heating mechanism. Most researchers agreed that chondrules had formed as closed systems without losses to, or gains from, the nebula (see [6] for references), hence requiring a flash heating mechanism in order to avoid interactions with the nebular gas.

10

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

According to this view, chondrule formation had little effect on the chondritic matter aside from textural, and therefore no connection to the chemical and isotopic differences between CIs and other chondrite classes summarized above. As discussed in the following sections, however, a growing number of researchers are challenging this view and arguing for open system behavior based on different lines of evidence. 4.2. Forming chondrules from microdroplets and dust Wood [6] considered that creating a population of dustballs in the first place was far from straightforward even if this had been commonly ignored. Although this problem does no longer appear critical [50,51], the alternative favored by Wood [6] remains useful to explain the peculiar textures of chondrules in carbonaceous chondrites as the one in Fig. 3b. To account for these, Wood [6] suggested forming chondrules from an accumulation of hot microdroplets and dust, although he recognized that one of the problems with this mechanism is the scarcity of preserved microchondrules in the matrices of primitive chondrites. These might however have been lost and accumulated elsewhere, for example in CH chondrites, as these consist almost exclusively of microchondrules and debris with an average size of 20 Am. 4.3. Forming chondrules by condensing liquid Condensation of liquids from an enriched gas constitutes another alternative that has been receiving growing attention. Ebel and Grossman [52] showed that liquids with compositions close to those of chondrules could be stable in local environments with high total pressures or high partial pressures of the relevant elements. This hypothesis provides a simple explanation to the absence of Rayleigh distillation induced isotopic fractionations in chondrules. In addition, Krot [53] argues that the extreme variety of compositions of chondrules in CHs (which range from highly refractory-enriched to highly refractory-depleted) can only be ascribed to fractional condensation. It is interesting to note that this hypothesis is not exclusive of the microdroplet mechanism favored by Wood [6] as the condensation might even have been to microdroplets which subsequently would have agglomerated, as he suggested [6].

4.4. Forming chondrules from earlier generations of chondrules Yet another possibility advocated by Alexander [54] is that chondrules were made from the recycling of earlier generations of chondrules, based on their chemical compositions and on the evidence for successive episodes of melting. This hypothesis is particularly attractive for chondrules in ordinary chondrites in which the least melted chondrules (presumed to be the least transformed and hence the closest to their precursors [55]) are aggregates of fine crystals with occasional coarser relicts, rather than looking as if they coalesced from microdroplets, as in carbonaceous chondrites [55 – 57]. Ordinary chondrites could thus contain mostly chondrules made from recycled precursors whereas carbonaceous chondrites would contain more chondrules of the earliest generations. 4.5. Was the system open or closed? Chondrule compositions are a key issue as they range from refractory-rich to volatile-rich. This variety of compositions could either be inherited from the precursors or result from the chondrule-forming event, depending on whether the system was closed or open to the surrounding gas, or derive from a combination of both. In the canonical nebular models, where condensates formed dustballs as temperature fell, the system is closed: volatile-poor chondrules were made from volatile-poor material, while volatile-rich chondrules require a flash heating mechanism to preserve their volatile content. The experiments of Hewins et al. [58] and Yu et al. [21] showed, however, that type II chondrule bulk compositions are impossible to explain with melting in a canonical nebula, even with flash heating, because retaining volatiles such as S, Na or K requires total pressures and oxygen fugacities, or partial pressures of these elements, several orders of magnitude above the nebular values. These partial pressures are not known, so the time scale for the formation of natural chondrules cannot be specified, but they ought to have been high enough to allow recondensation of volatiles into chondrules. This process may have contributed to the apparent lack of isotopic fractionation induced by evaporation during chondrule formation (e.g., [9,10]).

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

Comparing chondrule chemistry with their textures also points to an open system behavior; Hewins et al. [55] found an inverse correlation between the grain size of ordinary chondrite type I chondrules and their volatile content. The finest (least melted) chondrules have Mg-normalized bulk compositions close to that of CI chondrite and contain no metal but abundant Nibearing Fe-sulfide [56]. All the intermediates exist between these fine-grained chondrules and the typical volatile-depleted and metal-bearing coarser-grained ones. This indicates that type I chondrules may have started with unaltered CI compositions and evolved to their present compositions during melting, as was experimentally achieved by Cohen [23]. In such a case, most chondritic metal might be a product from chondrules (made primarily by S loss [56], although some carbon-reduction might also occur in chondrule formation). This appears consistent with the present mineralogy of chondrule-free CI chondrites which are metal free (but contain Ni-bearing Fe-sulfide and magnetite instead) and may have done so even before they were altered on their parent bodies. 4.6. Geochemical consequences of open system behavior By producing metallic melts which have the ability to physically separate from silicate melts (e.g., [59]) and by losing volatiles which will only partly recondense (e.g., [37]), chondrules thus have the ability to generate both the metal/silicate and the refractory/ volatile fractionations that characterize the chondrite classes (Figs. 7 and 8). In that framework, the chemical complementarity between matrix and chondrules [35 – 37] finds a natural explanation as chondrules and matrix would both have evolved from unaltered CI type material, one gaining part of the material (volatiles and metal) lost by the other. There is also some evidence of condensation into chondrule rim material. The present oxygen isotopic signatures of chondrules are still poorly understood and those of their precursors are not known. Luck et al. [41] discuss in detail the origin of the correlations between bulk 63Cu and 16O excesses in carbonaceous and in ordinary chondrites together with Al/Mn ratios. These authors suggest mixing an 16O enriched and a copper-poor but 63 Cu enriched refractory component together with another component close to CI in 16O and 63Cu or

11

slightly more depleted, possibly the nebular gas. Such interaction could clearly have taken place during chondrule formation as Yu et al. [20] experimentally demonstrated that an exchange with the ambient gas could modify the oxygen signatures of chondrules, as Varley et al. [60] found a mild correlation between D17O and the extent of chondrule melting in CR chondrites and as FeO-rich chondrules in CRs [61] and CMs [62] appear more depleted in 16O than their FeO-poor neighbours. This model might explain the oxygen isotopic signatures of ordinary chondrite chondrules which would have derived from unaltered CI type precursor material. Carbonaceous chondrite chondrules however require the addition of 16O and 63 Cu, presumably in an enriched refractory component as proposed by Luck et al. [41]. Alternative explanations for the mass independent fractionation of oxygen in meteorites, involving symmetry-dependent chemical reactions (e.g., [63]) or CO self shielding [64] or linked to Galactic chemical evolution [65], might allow us to derive chondrules from both ordinary and carbonaceous chondrites from unaltered CI material. It is however unclear that they would also provide an explanation for the correlations observed between bulk 63Cu and bulk 16O excesses in carbonaceous and ordinary chondrites and between 63Cu excesses and Al/Mn ratios. In addition, the presence of a refractory component in the precursors of chondrules from CO and CV chondrites but not from ordinary chondrites, (as shown by [7] based on REE signatures) yields support to the previous model. 4.7. The relationship between chondrules and CAIs This raises the issue of the relationship between chondrules and CAIs, the most likely candidates for the refractory component of chondrule precursors in carbonaceous chondrites, which happen to be very abundant (f 10– 13%) in CV and CO chondrites and very rare ( < 1%?) in ordinary chondrites (e.g., [66]). CAIs are the oldest objects of the solar system and appear to have been formed 2.5 F 1.2 Ma before chondrules [12]. Many of them have also experienced melting by a process that is currently unknown and that it is tempting to relate to that which later formed chondrules. It is however unlikely that there was a gradual change with time from CAI formation

12

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

to chondrule formation as the bulk of the two populations appears distinct both in age and in composition. A few intermediate objects however do exist such as Al-rich or forsterite –fassaite [30] chondrules. These might have been formed from precursors containing excess CAI material. The presence of CAI material within chondrule precursors poses another significant problem discussed in the next section: how could these objects have been prevented for more than 2 million years from drifting into the Sun [6]?

5. Where, how and when? The wide variety of proposed chondrule-forming mechanisms (e.g., Fig. 1) fall into two broad categories depending on whether they are assumed to have operated in the vicinity of newly formed planets or within the protoplanetary nebula. Following the publication 30 years ago of a very influential paper [67], a consensus seemed to have been reached by most scientists in the field (mostly petrologists at the time) that chondrules could only have been formed in a nebular setting, although a few dissenting voices always existed (e.g., [47]). The pendulum recently started swinging the other way, mostly under the influence of isotope geochemists who inferred, based on extinct radioactive isotopes such as 26Al [68,69] and 53Mn [70], that the formation of chondrules lasted from < 1 up to 4 million years after CAIs. 5.1. Forming chondrules in a planetary environment Assuming these ages are crystallization (rather than metamorphic) ages, forming chondrules over such an extended period raises the problem of preventing the oldest ones and the CAIs from drifting into the growing Sun so that they are able to mix with younger ones. While the chronological interpretation of extinct radionuclides alone could be questioned (as they might have been initially heterogeneously distributed or secondarily disturbed by parent-body processes), it was recently beautifully confirmed by the Pb –Pb data of Amelin et al. [12]. According to this interpretation, the youngest chondrules [68] have roughly the same age as the oldest phosphate age for a metamorphosed chondrite, 4563 Ma [48].

If the environment in which CAIs and chondrules formed did not change significantly over time, then protoplanets such as those which eventually became differentiated already existed soon after the oldest CAIs were melted and they provide one easy way to store these objects and prevent them from drifting into the sun. Additional benefits of planetary models are that they explain the presence of lithic clasts within chondrites [47] and the textures of chondrites as seen in Fig. 3, and especially they provide an atmosphere for melting volatile-rich chondrules much more suitable than the open solar nebula. They do not, however, allow us to generate the chemical fractionation between chondrite classes easily in the chondruleforming event, as seems to be suggested from the comparison between CIs and other chondrites. The possibilities for generating chondrules in a planetary environment are numerous. Most of them involve collisions and will not be reviewed here as their physics needs to be investigated in detail, as was recently done for nebular models. A recent hybrid planetary-nebular model [71] produces chondrules by heating nebular particles or debris of first generation planetesimals by bow-shocks around eccentric planetesimals driven by Jovian resonances. 5.2. Forming chondrules in a nebular environment The recent revisiting of planetary models seems timely as our knowledge of the time scales and processes at work in a disk containing colliding planetary embryos is evolving. It however does not appear fully justified based only on the problem of preventing older generations of CAIs and chondrules from drifting into the Sun, because many other solutions to this problem exist. One involves a first step of accretion immediately following chondrule formation to form bodies which are later disrupted and reassembled, mixing CAIs and chondrules of different ages. In a purely nebular context, outward radial diffusion in a weakly turbulent nebula will preserve CAIs [72]. The density of the gas may also be low enough that the residence time of CAIs and chondrules reaches several million years and, even if the bulk of the CAIs and chondrules spiral into the Sun, the tail of the distribution will still diffuse upstream [73].

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

Aside from condensation of a liquid from the nebular gas (e.g., [52]), models to form chondrules in a nebular setting are even more varied than for planetary environments. The present review concentrates on the two that were developed or significantly improved and received much attention since the summary sketched in Fig. 1. The model of Shu et al. [2] involves the inner edge of the protoplanetary disk where matter is partly accreted into the Sun and partly ejected by bipolar outflows. Heating by the Sun and solar flares melts dustballs brought to the inner edge of the disk makes CAIs and chondrules. Some of these are subsequently thrown back by an ‘‘X-wind’’ onto the disk in the asteroidal belt region. This model explains the narrow size distribution of chondritic components in each chondrite class (Fig. 5) by sorting and also allows generating some of the short-lived radioactive nuclides recently detected in chondrules [15] and in CAIs (e.g., [74,75]) by irradiation. However, it fails to explain the cooling rates observed for chondrules, the generation of matrix with a different composition in different locations (to explain the apparent matrix –chondrule complementarity) and the provision of a mechanism for storing CAIs and chondrules over extended periods (to explain their Pb –Pb age difference). The nebular shock model of Desch and Connolly [19] and of Ciesla and Hood [76] does better in some of these respects. It involves melting chondrules (and possibly also CAIs) as the result of the propagation of shock waves within the nebula and, by taking into account the radiation absorbed by the chondrules both from the heated gas and from the neighboring chondrules, generates cooling rates that match those estimated for chondrules [19]. Transient atmospheres are generated in clumps of evaporating dust and chondrules, allowing formation of volatile-rich chondrules. The matrix is formed from unevaporated dust and some volatiles lost from chondrules, so the complementarity between it and chondrules falls out of the model. Size sorting (see Fig. 5) of chondrites is achieved by concentrating chondrules between eddies in a turbulent nebula (after Cuzzi et al. [77]) and accreting separately chondrules with different aerodynamic properties (which depend on their sizes). Finally, this model [19] predicts a correlation between chondrule density in the source region, cooling

13

rate and temperature that generates relative abundances of the different chondrule textural types corresponding to those observed in chondrites. This represents a milestone in the present author’s opinion as it is the first model that closely matches some of the significant properties of chondrites. Like other nebular models, however, it needs to appeal to a mechanism such as [72] to prevent CAIs and chondrules from being dragged by the nebular gas into the growing Sun. 5.3. The possible consequences of chondrule formation in a nebular setting The stopping time of chondrules due to gas drag depends on their radius which correlates with their FeO content, so type I and II chondrules are expected to separate [78] and so are metallic particles which are smaller even than type I chondrules. By allowing such sorting, chondrule formation in a nebular setting thus has the ability to generate the various chondrite classes. This possibility is especially attractive for ordinary chondrites in which average chondrule diameters range from 300 Am (in H) to 700 Am (in L) to 900 Am (in LL) [66] while the vol.% of chondrule material present as type I decreases from 57% (in H) to 40% (in L) to 25% (in LL) [79] and the abundance of metal (which has aerodynamic properties closer to those of type I chondrules [80]) decreases in parallel. This model even holds a potential explanation for the oxygen isotopic signatures of the three classes, as there is no difference between randomly chosen individual chondrules of each class (which span the whole range of values of the bulk samples from the three classes) but two reasons make it likely that volatile-rich type II (concentrated in LLs) are on average slightly more depleted in 16O than volatilepoor type I: (1) Clayton et al. [81] showed that in ordinary chondrites the larger chondrules tend to be more depleted in 16O; (2) type II chondrules in carbonaceous chondrites tend to be more depleted in 16O than their type I counterparts [61,62]. Note that this model might even be extended to the case of the Earth if it was made mostly from type I chondrules as suggested by Hewins and Herzberg [1], as a higher content of type I chondrules than in H chondrites would naturally place it at the intercept

14

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

of the Y + R and of the terrestrial fractionation line in Fig. 9. Chondrule formation in a nebular setting may also have played a role in the earliest stages of accretion by sticking together particles while they are still partly melted or above the glass transition temperature as the textures in Fig. 3 suggest. Much of the earliest accretion may however be due to electrostatic attraction [50,51].

seems to be suggested by the differences between carbonaceous and ordinary chondrites discussed throughout this paper and shown in Fig. 3. The study of new chondrite classes or (in the distant future) of samples returned from chondritic asteroids might help in that respect.

6. Future directions

Until fairly recently, the solar nebula has been treated as a ‘‘black box’’ in which any poorly understood effect could have happened and there has in the past been too much distance between meteoriticists and astrophysicists. Chondrule formation however needs to be linked to realistic astrophysical constraints. The 1994 ‘‘Chondrule and the Protoplanetary Disk Conference’’ followed by a book [82], started bridging that gap. It will hopefully be narrowed further with new chondrule formation models, which, after [19], will focus on making detailed petrographic predictions.

Despite the significant recent advances described above (such as obtaining the first absolute ages of chondrules), major problems still plague all the possible chondrule-forming mechanisms that have been proposed up to now and we remain unable to unequivocally decide between the two types of possible environments in which chondrules were formed. Progress may come from several directions: 6.1. More detailed observations of chondrules Precise dating of individual chondrules has only just begun and is sure to contribute significantly to our knowledge of the chondrule forming mechanism. Another new direction will involve a much better understanding of the details of the physical mechanisms involved within a melting and cooling chondrule. This will be achieved with the help of in situ isotopic and trace measurements relevant to exchanges between the solid, liquid and gaseous phases involved, coupled with experimental simulations. 6.2. More detailed observations of chondrites We are still unable to fully disentangle the effects of the chondrule-forming event from those of aqueous alteration and thermal metamorphism. These secondary effects may distort our present vision of chondrule precursors and of the chondrule-forming mechanism. Systematic studies of chondrule properties (such as their age, chemistry, textures, diameters) in a given meteorite and within a chondrite class will also provide tighter constraints. It is possible that all chondrules do not share the same type of precursors and/or formation mechanism as

6.3. More detailed observation and modeling of young planetary systems trying to better reproduce chondrule/chondrite properties

Acknowledgements Like at least one previous Frontiers author [83], I want to thank Alex Halliday for ‘‘early encouragement and recent patience’’. I also want to thank for their insight my colleagues from MNHN: M. BourotDenise, C. Perron and F. Robert and from Rutgers University: B. Cohen and Y. Yu. P. Cassen and S. Desch made helpful suggestions concerning the astrophysical aspects, D. Ben Othman and J.-M. Luck concerning Cu isotopes and H. Palme kindly let me copy Fig. 7 from one of his manuscripts. He, S. Desch and C. M.O’D Alexander are thanked for reviews which improved this manuscript. Last but not least, I thank R. Hewins for scientific and material support throughout this work. [AH]

References [1] R.H. Hewins, C.T. Herzberg, Nebular turbulence, chondrule formation, and the composition of the earth, Earth Planet. Sci. Lett. 144 (1996) 1 – 7.

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17 [2] F. Shu, H. Shang, A.E. Glassgold, T. Lee, X-rays and fluctuating x-winds from protostars, Science 277 (1997) 1475 – 1479. [3] J.A. Wood, Chondrites: Tight-Lipped Witnesses to the Beginning, Harold Masursky Plenary Lecture, LPSC, Houston, 2000. [4] R.A. Kerr, A meteoriticist speaks out, his rocks remain mute, Science 293 (2001) 1581 – 1584. [5] R.H. Hewins, Chondrules and the protoplanetary disk: an overview, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, pp. 3 – 9. [6] J.A. Wood, Unresolved issues in the formation of chondrules and chondrites, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, pp. 55 – 69. [7] K. Misawa, N. Nakamura, Origin of refractory precursor components of chondrules from carbonaceous chondrites, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, pp. 99 – 105. [8] H. Yurimoto, J.T. Wasson, Extremely rapid cooling of a carbonaceous-chondrite chondrule containing very 16O-rich olivine and a 26Mg excess, Geochim. Cosmochim. Acta 66 (2002) 4355 – 4363. [9] C.M.O’D. Alexander, J.N. Grossman, J. Wang, B. Zanda, M. Bourot-Denise, R.H. Hewins, The lack of potassium isotopic fractionation in Bishunpur chondrules, Meteorit. Planet. Sci. 35 (2000) 859 – 868. [10] C.M.O’D. Alexander, J. Wang, Iron isotopes in chondrules: implications for the role of evaporation during chondrule formation, Meteorit. Planet. Sci. 36 (2001) 419 – 428. [11] T.D. Swindle, A.M. Davis, C.M. Hohenberg, G.J. MacPherson, L.E. Nyquist, Formation times of chondrules and Ca – Al-rich inclusions: constraints from short-lived radionuclides, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, pp. 77 – 86. [12] Y. Amelin, A.N. Krot, I.D. Hutcheon, A.A. Ulyanov, Lead isotopic ages of chondrules and calcium – aluminum-rich inclusions, Science 297 (2002) 1678 – 1683. [13] A.N. Halliday, The Origin and Earliest History of the Earth, in: Treatise of Geochemistry, vol. 1, Elsevier, Amsterdam, 2004. [14] S.J. Kortenkamp, E. Kokubo, S.J. Weidenschilling, Formation of planetary embryos, in: R.M. Canup, K. Righter (Eds.), Origin of the Earth and Moon, University of Arizona Press, Tucson, 2000, pp. 85 – 100. [15] F. Robert, M. Chaussidon, Boron and lithium isotopic composition in chondrules from the Mokoı¨a meteorite, Lunar Planet. Sci. XXXIV (2003) 1344, (pdf.). [16] S.J. Desch, H.C. Connolly, G. Srinivasan, An interstellar origin for the beryllium 10 in calcium-rich, aluminum-rich inclusions, Astrophys. J. 602 (2004) 528 – 542. [17] P.M. Radomsky, R.H. Hewins, Formation conditions of pyroxene – olivine and magnesian olivine chondrules, Geochim. Cosmochim. Acta 54 (1990) 3475 – 3490. [18] Y. Yu, R.H. Hewins, Transient heating and chondrule formation-evidence from Na loss in flash heating simulation experiments, Geochim. Cosmochim. Acta 62 (1998) 159 – 172.

15

[19] S.J. Desch, H.C. Connolly Jr., A model for the thermal processing of particles in solar nebula shocks: application to cooling rates of chondrules, Meteorit. Planet. Sci. 37 2002, pp. 183 – 208. [20] Y. Yu, R.H. Hewins, R.N. Clayton, T.K. Mayeda, Experimental study of high temperature oxygen isotope exchange during chondrule formation, Geochim. Cosmochim. Acta 59 (1995) 2095 – 2104. [21] Y. Yu, R.H. Hewins, C.M.O’D. Alexander, J. Wang, Experimental study of evaporation and isotopic mass fractionation of potassium in silicate melts, Geochim. Cosmochim. Acta 67 (2003) 773 – 786. [22] B.A. Cohen, R.H. Hewins, Y. Yu, Evaporation in the young solar nebula as the origin of ‘‘just-right’’ melting of chondrules, Nature 406 (2000) 600 – 602. [23] B.A. Cohen, Chondrule formation by open system melting of nebular condensates, PhD thesis, Rutgers University (2002). [24] C.M.O’D. Alexander, Chemical equilibrium and kinetic constraints for chondrule and CAI formation, Geochim. Cosmochim. Acta (in press). [25] A.J. Brearley, Nature of matrix in unequilibrated chondrites and its possible relationship to chondrules, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, pp. 137 – 151. [26] D.W.G. Sears, R.T. Dodd, Overview and classification of meteorites, in: J.F. Kerridge, M.S. Matthews (Eds.), Meteorites and the Early Solar System, University of Arizona Press, Tucson, 1988, pp. 3 – 31. [27] M. Rotaru, J.-L. Birck, C.J. Alle`gre, Clues to early solar-system history from chromium isotopes in carbonaceous chondrites, Nature 358 (6386) (1992) 465 – 470. [28] N. Dauphas, B. Marty, L. Reisberg, Molybdenum nucleosynthetic dichotomy revealed in primitive meteorites, Astrophys. J. 569 (2002) L139 – L142. [29] L.A. Leshin, A.E. Rubin, K.D. McKeegan, The oxygen isotopic composition of olivine and pyroxene from CI chondrites, Geochim. Cosmochim. Acta 61 (1997) 835 – 845. [30] B. Zanda, G. Libourel, Ph. Blanc, Source chondrules for refractory forsterites in primitive chondrites, revision for, Meteorit. Planet. Sci. (2004), in press. [31] F. Robert, D. Gautier, B. Dubrulle, The Solar system D/H ratio: observations and theories, ISSI: Dust to Terrestrial Planets, Berne, Kluwer, Dordrecht, 2000, pp. 201 – 224. [32] E. Deloule, F. Robert, J.C. Doukhan, Interstellar hydroxyl in meteoritic chondrules: implications of water in the inner solar system, Geochim. Cosmochim. Acta 62 (1998) 3367 – 3378. [33] H. Palme, Chemical and isotopic heterogeneity in protosolar matter, Philos. Trans. R. Soc. London Ser., A 359 (2001) 2061 – 2075. [34] J.W. Larimer, E. Anders, Chemical fractionation in meteorites: II. Abundance patterns and their interpretation, Geochim. Cosmochim. Acta 31 (1967) 1239 – 1270. [35] J.A. Wood, Meteoritic constraints on processes in the solar nebula, in: D.C. Black, M.S. Matthews (Eds.), Protostars and Planets II, University of Arizona Press, Tucson, 1985, pp. 687 – 702.

16

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17

[36] H. Palme, S. Klerner, Formation of chondrules and matrix in carbonaceous chondrites, Meteorit. Planet. Sci. 35 (2000) A124. [37] B. Zanda, M. Humayun, R.H. Hewins, M. Bourot-Denise, A.J. Campbell, The Relationship Between Volatile Element Patterns and Chondrule Textures in CRs and OCs, Goldschmidt Conference Abtracts, 2002. [38] R. N. Clayton, N. Onuma, L. Grossman, T.K. Mayeda, Distribution of the pre-solar component in Allende and other carbonaceous chondrites, Earth Planet. Sci. Lett. 34 (1977) 209 – 224. [39] R.N. Clayton, T.K. Mayeda, The oxygen isotope record in Murchison and other carbonaceous chondrites, Earth Planet. Sci. Lett. 67 (1984) 151 – 161. [40] E.D. Young, S.S. Russell, Oxygen reservoirs in the early solar nebula inferred from an Allende CAI, Science 282 (1998) 452 – 455. [41] J.-M. Luck, D. Ben Othman, J.-A. Barrat, F. Albare`de, Coupled 63Cu and 16O excesses in chondrites, Geochim. Cosmochim. Acta 67 (2003) 143 – 151. [42] R.N. Clayton, T.K. Mayeda, Oxygen isotope classification of carbonaceous chondrites, Lunar Planet. Sci. XX (1989) 169 – 170. [43] R.N. Clayton, Oxygen isotopes in meteorites, Annu. Rev. Earth Planet. Sci. 21 (1993) 115 – 149. [44] G.W. Kallemeyn, A.E. Rubin, J.T. Wasson, The compositional classification of chondrites: VII. The R chondrite group, Geochim. Cosmochim. Acta 60 (1996) 2243 – 2256. [45] Ph. Bland, personal communication (2004). [46] D. Ben Othman, J.-M. Luck, personal communication (2004). [47] R. Hutchison, Chondrules and their associates in ordinary chondrites: a planetary connection? in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, 1996, pp. 311 – 318, Cambridge. [48] C. Go¨pel, G. Manhe`s, C.J. Alle`gre, U – Pb systematics of phosphates from equilibrated ordinary chondrites, Earth Planet. Sci. Lett. 121 (1994) 153 – 171. [49] K. Misawa, A. Yamaguchi, Zircons in eucrites: occurrence, possible origin and U – Pb isotopic systematics, Lunar Planet. Sci. XXXII (2001) 1676. [50] N. Haghighipour, Growth of dust particles and accumulation of centimeter-sized objects in the vicinity of a pressure-enhanced region of a solar nebula, Lunar Planet. Sci. XXXV (2004) 2001. [51] S.G. Love, D.R. Pettit, Fast, repeatable clumping of solid particles in microgravity, Lunar Planet. Sci. XXXV (2004) 1119. [52] D.S. Ebel, L. Grossman, Condensation in dust-enriched systems, Geochim. Cosmochim. Acta 64 (2000) 339 – 366. [53] A.N. Krot, et al, Ferrous silicate spherules with euhedral iron – nickel metal grains from CH carbonaceous chondrites: evidence for supercooling and condensation under oxidizing conditions, Meteorit. Planet. Sci. 35 (2000) 1249 – 1258. [54] C.M.O’.D. Alexander, Recycling and volatile loss in chondrule formation, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, pp. 233 – 241. [55] R.H. Hewins, Y. Yu, B. Zanda, M. Bourot-Denise, Do nebular fractionations, evaporative losses, or both, influence chondrule compositions? Antarct. Meteor. Res. 10 (1997) 294 – 317.

[56] B. Zanda, M. Bourot-Denise, R.H. Hewins, Chondrule precursors: the nature of the S- and Ni-bearing phase(s), Lunar Planet. Sci. XXVII (1996) 1485 – 1486. [57] J.W. Nettles, G.E. Lofgren, H.Y. McSween Jr., Recycled ‘‘chondroids’’ in LEW86018: a petrographic study of chondrule precursors, Lunar Planet. Sci. XXXIII (2002) 1752. [58] R.H. Hewins, B. Zanda, Y. Yu, M. Bourot-Denise, Towards a new model for chondrules, Paul Pellas Symposium abstracts, Muse´um National d’Histoire Naturelle, Paris, 1998, pp. 31 – 33. [59] G.K. Benedix, T.J. McCoy, T.L. Dickinson, G.E. Lofgren, Partial melting of chips of the indarch (EH4) meteorite: further insights into melt migration, Meteorit. Planet. Sci. 36 (2001) A18, (Suppl.). [60] L.R. Varley, L.A. Leshin, Y. Guan, B. Zanda, M. BourotDenise, Oxygen isotopic composition of Renazzo chondrule olivine and comparison with extent of chondrule melting, Lunar Planet. Sci. XXXIV (2003) 1899. [61] H.C. Connolly Jr., M.K. Weisberg, G.R. Huss, On the nature and origins of FeO-rich chondrules in CR2 chondrites: a preliminary report, Lunar Planet. Sci. XXXIV (2003) 1770. [62] I. Jabeen, H. Hiyagon, Oxygen isotopes in isolated and chondrule olivines of Murchison, Lunar Planet. Sci. XXXIV (2003) 1551. [63] M.H. Thiemens, Mass-independent isotopic effects in chondrites: the role of chemical processes, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, 1996, pp. 107 – 118, Cambridge. [64] R.N. Clayton, Solar system: self-shielding in the solar nebula, Nature 415 (2002) 860 – 861. [65] D.D. Clayton, Isotopic anomalies: chemical memory of galactic evolution, Astrophys. J. 334 (1988) 191 – 195. [66] A.J. Brearley, R.H. Jones, Chondritic meteorites, in: J.J. Papike (Ed.), Planetary Materials, Reviews in Mineralogy vol. 36, 1998, 398 pp. [67] G.K. Taylor, E.R.D. Scott, K. Keil, Cosmic setting for chondrule formation, in: E.A. King (Ed.), Chondrules and Their Origins, Lunar and Planetary Institute, Houston, 1983, pp. 262 – 278. [68] S.S. Russell, G.R. Huss, G.J. MacPherson, G.J. Wasserburg, Early and late chondrule formation: new constraints for solar nebula chronology from 26Al/27Al in unequilibrated ordinary chondrites, Lunar Planet. Sci. XXVIII (1997) 1209 – 1210. [69] S. Mostefaoui, N.T. Kita, S. Togashi, S. Tachibana, H. Nagahara, Y. Morishita, The relative formation ages of ferromagnesian chondrules inferred from their initial aluminum26/aluminum-27 ratios, Meteorit. Planet. Sci. 37 (2002) 421 – 438. [70] G.W. Lugmair, A. Shukolyukov, Early solar system events and timescales, Meteorit. Planet. Sci. 36 (2001) 1017 – 1026. [71] S.J. Weidenschilling, F. Marzari, L.L. Hood, The origin of chondrules at Jovian resonances, Science 279 (1998) 681 – 684. [72] J.N. Cuzzi, S.S. Davis, A.R. Dobrovolskis, Blowing in the wind: II. Creation and redistribution of refractory inclusions in a turbulent protoplanetary nebula, Icarus 166 (2) (2003) 385 – 402. [73] S. Desch, Personal communication (2003).

B. Zanda / Earth and Planetary Science Letters 224 (2004) 1–17 [74] M. Chaussidon, F. Robert, K.D. McKeegan, S. Krot, Li, Be, B distribution and isotopic composition in refractory inclusions from primitive chondrites: a record of irradiation processes in the protosolar nebula, Meteoritics 36 2001, pp. A40, (suppl.). [75] M. Chaussidon, F. Robert, K.D. McKeegan, Incorporation of short-lived 7Be in one CAI from the allende meteorit, Lunar Planet. Sci. XXXIII (2002) 1563. [76] F.J. Ciesla, L.L. Hood, The nebular shock wave model for chondrule formation: shock processing in a particle – gas suspension, Icarus 158, 2002, pp. 281 – 293. [77] J.N. Cuzzi, R.C. Hogan, J.M. Paque, A.R. Dorbovoskis, Sizeselective concentration of chondrules and other small particles in protoplanetary nebula turbulence, Astrophys. J. 546 (2001) 496 – 508. [78] J.N. Cuzzi, A.R. Dobrovolskis, R.C. Hogan, Turbulence, chondrules, and planetesimals, in: R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge Univ. Press, Cambridge, 1996, pp. 35 – 43. [79] M. Bourot-Denise, unpublished data. [80] K.E. Kuebler, H.Y. McSween, W.D. Carlson, D. Hirsch, Sizes

17

and masses of chondrules and metal-troilite grains in ordinary chondrites: possible implications for nebular sorting, Icarus 141 (1999) 96 – 106. [81] R.N. Clayton, T.K. Mayeda, J.N. Goswami, E.J. Olsen, Oxygen isotopic studies of ordinary chondrites, Geochim. Cosmochim. Acta 55 (1991) 2317 – 2337. [82] R.H. Hewins, R.H. Jones, E.R.D. Scott (Eds.), Chondrules and the Protoplanetary Disk, Cambridge University Press, Cambridge, 1996, 346 pp. [83] A.D. Anbar, Iron stable isotopes: beyond biosignatures, Earth Planet. Sci. Lett. 217 (2004) 223 – 236. Brigitte Zanda is currently the curator in charge of the meteorite collection at the Paris Muse´um. Most of her recent work has involved petrologic studies of chondrules and chondrites, coupled with collaborations with geochemists and experimentalists. Her goal is to investigate the role of chondrule formation in the genesis of protoplanets.